Series battery pack health state estimation method based on open-circuit voltage reconstruction

Abstract

The invention discloses a series battery pack health state estimation method based on open-circuit voltage reconstruction, and the method comprises the following steps: collecting the voltage and current of each single battery in a battery pack and the overall environment temperature data of the battery pack, and carrying out the sliding window processing of the noise-reduced time sequence data; aiming at charging electric quantity data under different temperatures and charging multiplying powers, correcting offset through a temperature compensation function, and establishing a characteristic sequence based on the corrected data; by synchronously quantifying the charge state difference and the capacity difference of the single batteries, a series battery pack capacity estimation model is established, and the technical scheme of the invention has wider applicability, can adaptively capture and predict the capacity performance change of the battery pack, and improves the capacity estimation accuracy of the battery pack. A series battery pack capacity estimation model is established through charging data of each single battery of the battery pack, optimal parameters are determined by minimizing an objective function, and the health condition of the battery pack can be calculated according to the optimal parameters of each battery.

Description

Technical Field

[0001] This invention relates to the field of energy storage battery technology, specifically to a method for estimating the health status of a series battery pack based on open-circuit voltage reconstruction. Background Technology

[0002] Existing methods for estimating the state of health (SOH) of series-connected battery packs typically rely on indirect assessments based on capacity decay or internal resistance changes, but these methods have significant limitations. First, the capacity differences and SOC imbalances among individual cells in a series-connected battery pack are difficult to quantify dynamically, leading to overall capacity estimation bias. During long-term cycling, differences in aging rates between cells can cause SOC distribution shifts, and traditional methods lack accurate modeling of these dynamic differences, failing to correct accumulated errors in real time. Second, temperature fluctuations and changes in charge / discharge rates significantly impact battery characteristics, but existing models lack effective environmental compensation mechanisms, making the estimation results susceptible to environmental interference, especially exacerbated by errors under high-temperature or high-rate charge / discharge conditions.

[0003] Furthermore, existing parameter optimization algorithms (such as particle swarm optimization and genetic algorithms) are prone to getting trapped in local optima during model solving, affecting model convergence and accuracy. These algorithms are sensitive to initial parameters and struggle to balance global search and local exploitation capabilities, leading to unstable capacity estimation results. These issues collectively limit the effectiveness of tandem battery pack health state estimation technology in complex real-world scenarios.

[0004] To address this, a health state estimation method for series battery packs based on open-circuit voltage reconstruction is proposed. Summary of the Invention

[0005] The purpose of this invention is to provide a health state estimation method for series battery packs based on open-circuit voltage reconstruction, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for estimating the health status of a series battery pack based on open-circuit voltage reconstruction, comprising the following steps:

[0007] S1. Collect the voltage and current of each individual cell in the battery pack and the overall ambient temperature data of the battery pack, and apply sliding window processing to the noise-reduced time-series data.

[0008] S2. For charging power data under different temperatures and charging rates, the offset is corrected by a temperature compensation function, and a feature sequence is established based on the corrected data.

[0009] S3. By synchronously quantifying the differences in state of charge and capacity of individual cells, a capacity estimation model for series battery packs is established, and the parameters are optimized based on the improved Seagull algorithm.

[0010] S4. Calculate the maximum usable capacity and health status of the series battery pack by combining the optimal parameters.

[0011] Preferably, in step S1, the data processing procedure includes:

[0012] S11. Perform mixed noise reduction processing on the original voltage and current signals, use wavelet transform to decompose high-frequency noise components, and suppress transient interference through Kalman filter;

[0013] S12. A degenerate feature sequence with temporal continuity is constructed using the sliding window method on the denoised data.

[0014] Preferably, in step S2, the temperature compensation function is:

[0015] Q Ah =∑γI(I / I ref ) α-1 (T ref / T) β Δt;

[0016] Among them, Q Ah To accumulate charging capacity, I ref For the reference current, T ref The reference temperature is γ, which is a coefficient in Ah, and α and β are coefficients for current and temperature, respectively.

[0017] Preferably, in step S3, the capacity estimation model for the series battery pack is established as follows:

[0018] S31, via Q start The difference in state of charge between cells is quantified to calculate and update the initial state of charge of each cell:

[0019]

[0020] Among them, SOC initial It is the initial SOC, Q start ΔQ represents the unusable capacity of the battery that was not charged when charging was stopped. max Q is the charging capacity from this moment until the charging ends. aging This is the actual capacity;

[0021] S32, combined with Q part The SOC change during the charging process is described, and its calculation process is as follows:

[0022]

[0023] Among them, SOC aging It is the actual SOC, Q part This is a partial charging capacity;

[0024] S33. The SOC scale of the aged battery is converted to that of the new battery to quantify the capacity difference, and the estimated value of the open-circuit voltage is solved using the OCV-SOC relationship of the new battery. The calculation process is as follows:

[0025]

[0026] Among them, SOC fresh It is the SOC, Q, scaled to the level of a fresh battery. fresh It is the capacity of a fresh battery. This is the estimated value of the open-circuit voltage at this time.

[0027] Preferred: The objective function for parameter solving is constructed using the root mean square error between the measured and estimated terminal voltages.

[0028]

[0029] θ=[Q start Q aging ,R];

[0030] Among them, U i R is the actual terminal voltage of the battery, and R is the polarization resistance during the charging process.

[0031] Preferably, in step S3, the specific process of parameter optimization includes:

[0032] S34. Initialize the seagull population size, spatial dimension, maximum number of iterations, and upper and lower bounds of the optimization parameters, where chaotic mapping is used to initialize the seagull population:

[0033] X n+1 =4X n (1-X n X∈[0,1];

[0034] S35. Each parameter updates its position according to the migration and predation behavior formulas and calculates a new fitness to obtain the optimal position. Variable A can be used to update the parameter position. The improved expression for variable A is as follows:

[0035]

[0036] In the seagull optimization algorithm, because seagulls have blind spots when moving, they may get trapped in local optima. Therefore, the seagull attack method has been updated:

[0037] T S (t)=C S (t)+M S (t);

[0038]

[0039] Among them, T S (t) represents the migratory position of seagulls adapted to multi-directional spirals, C S (t) is the latest position that does not conflict with the positions of other seagulls, M S (t) is the direction of the optimal solution, P bs (t) is the position of the optimal solution, x, y, z are the three dimensions of the current position, and a is the selection update threshold;

[0040] S36. Check if the current iteration count has reached the maximum value. If it has reached the set maximum iteration count, output the current best position; otherwise, continue the search.

[0041] Preferably, in step S4, calculating the maximum usable capacity and state of health of the series-connected battery pack includes:

[0042] S41. Solve for the optimal parameters of each cell in the battery pack based on minimizing the objective function, and calculate the maximum usable capacity of the series-connected battery pack using the optimal parameters:

[0043] Q pack =min(Q) aging，i -Q start，i );

[0044] S42. The health status of the battery pack is the ratio of the maximum usable capacity of the aged battery pack to that of the new battery pack.

[0045]

[0046] Compared with the prior art, the beneficial effects of the present invention are: the technical solution of the present invention has a wider applicability, can adaptively capture and predict the capacity performance changes of the battery pack, establish a series battery pack capacity estimation model through the charging data of each individual cell in the battery pack, determine the optimal parameters by minimizing the objective function, and calculate the health status of the battery pack based on the optimal parameters of each cell. Attached Figure Description

[0047] Figure 1 This is a flowchart of the health status estimation method for series battery packs of the present invention;

[0048] Figure 2 This is a schematic diagram showing the measured and estimated terminal voltage values ​​of the target battery under different aging conditions according to the present invention.

[0049] Figure 3 This is a schematic diagram illustrating the estimation error of the health status of the series battery pack of the present invention. Detailed Implementation

[0050] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0051] Please see Figure 1-3 This invention provides a technical solution: a method for estimating the health status of a series battery pack based on open-circuit voltage reconstruction, comprising the following steps:

[0052] S1. Collect the voltage and current of each individual cell in the battery pack and the overall ambient temperature data of the battery pack, and apply sliding window processing to the noise-reduced time-series data.

[0053] Furthermore, step S1 specifically includes:

[0054] S11. The original voltage and current signals are mixed and denoised. Wavelet transform is used to decompose the high-frequency noise components, and Kalman filter is used to suppress transient interference.

[0055] S12. Construct a time-continuous degenerate feature sequence using the sliding window method on the denoised data:

[0056]

[0057] Where N is the sampling point, U part I part These are voltage and current segments.

[0058] S2. For charging power data under different temperatures and charging rates, the offset is corrected by a temperature compensation function, and a feature sequence is established based on the corrected data.

[0059] Furthermore, in step S2, a charging quantity feature sequence is established based on the temperature compensation function. The specific steps are as follows:

[0060] S21. Since the battery has different usable capacities at different discharge rates and temperatures, a temperature compensation function is used to correct the accumulated charge during the charging process:

[0061] Q Ah =∑γI(I / I ref ) α-1 (T ref / T) β Δt;

[0062] Among them, Q Ah To accumulate charging capacity, I ref For the reference current, T ref The reference temperature is γ, the coefficient is in Ah, and α and β are the coefficients of current and temperature, respectively. Based on the experimental test results, the coefficients of the above function can be obtained by fitting using the least squares method.

[0063] S22. Generate a degradation feature sequence Q by combining the processed data. part =[Q1,Q2,…,QN Then, based on the input vector, a capacity estimation model for open-series battery packs is constructed.

[0064] S3. By synchronously quantifying the differences in state of charge and capacity of individual cells, a capacity estimation model for series battery packs is established, and the parameters are optimized based on the improved Seagull algorithm.

[0065] Furthermore, in step S3, a capacity estimation model for the series battery pack is constructed. The specific steps are as follows:

[0066] S31, via Q start The difference in state of charge between cells is quantified to calculate and update the initial state of charge of each cell:

[0067]

[0068] Among them, SOC initial It is the initial SOC, Q start ΔQ represents the unusable capacity of the battery that was not charged when charging was stopped. max Q is the charging capacity from this moment until the charging ends. aging That is the actual capacity.

[0069] S32, combined with Q part The SOC change during the charging process is described, and its calculation process is as follows:

[0070]

[0071] Among them, SOC aging It is the actual SOC, Q part This is a partial charging capacity.

[0072] S33. The SOC scale of the aged battery is converted to that of the new battery to quantify the capacity difference, and the estimated value of the open-circuit voltage is solved using the OCV-SOC relationship of the new battery. The calculation process is as follows:

[0073]

[0074] Among them, SOC fresh It is the SOC, Q, scaled to the level of a fresh battery. fresh It is the capacity of a fresh battery. This is the estimated value of the open-circuit voltage at this time.

[0075] Furthermore, the parameter optimization process based on the improved Seagull algorithm in step S3 is as follows:

[0076] S34. Initialize the seagull population size, spatial dimension, maximum number of iterations, and upper and lower bounds of the optimization parameters, where chaotic mapping is used to initialize the seagull population:

[0077] Xn+1 =4X n (1-X n ), X∈[0,1].

[0078] S35. Each parameter updates its position according to the migration and predation behavior formulas and calculates a new fitness to obtain the optimal position. Variable A can be used to update the parameter position. The improved expression for variable A is as follows:

[0079]

[0080] In the seagull optimization algorithm, because seagulls have blind spots when moving, they may get trapped in local optima. Therefore, the seagull attack method has been updated:

[0081] T S (t)=C S (t)+M S (t);

[0082]

[0083] Among them, T S (t) represents the migratory position of seagulls adapted to multi-directional spirals, C S (t) is the latest position that does not conflict with the positions of other seagulls, M S (t) is the direction of the optimal solution, P bs (t) is the position of the optimal solution, x, y, z are the three dimensions of the current position, and a is the selection and update threshold.

[0084] S36. Check if the current iteration count has reached the maximum value. If it has reached the set maximum iteration count, output the current best position; otherwise, continue the search.

[0085] Specifically, it also includes:

[0086] The objective function for parameter solving is constructed using the root mean square error between the measured and estimated terminal voltages.

[0087]

[0088] θ=[Q start Q aging ,R];

[0089] Among them, U i R is the actual terminal voltage of the battery, and R is the polarization resistance during the charging process.

[0090] like Figure 2 As shown, Figure 2The measured and estimated values ​​of the terminal voltage of the target battery under different aging conditions are shown. As the battery capacity decreases, the reconstruction error gradually increases, but the error is still within an acceptable range.

[0091] S4. Calculate the maximum usable capacity and health status of the series battery pack by combining the optimal parameters.

[0092] Furthermore, the calculation of the maximum usable capacity and health status of the series-connected battery pack in step S4 is as follows:

[0093] S41. Solve for the optimal parameters of each cell in the battery pack based on minimizing the objective function, and calculate the maximum usable capacity of the series-connected battery pack using the optimal parameters:

[0094] Q pack =min(Q) aging，i -Q start，i ).

[0095] S42. The health status of the battery pack is the ratio of the maximum usable capacity of the aged battery pack to that of the new battery pack.

[0096]

[0097] like Figure 3 As shown, Figure 3 The relative error in estimating the health status of the target series battery pack can be controlled within 3%.

[0098] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for estimating the health status of a series battery pack based on open-circuit voltage reconstruction, characterized in that, Includes the following steps: S1. Collect the voltage and current of each individual cell in the battery pack and the overall ambient temperature data of the battery pack, and apply sliding window processing to the noise-reduced time-series data. S2. For charging power data under different temperatures and charging rates, the offset is corrected by a temperature compensation function, and a feature sequence is established based on the corrected data. S3. By synchronously quantifying the differences in state of charge and capacity of individual cells, a capacity estimation model for series battery packs is established, and the parameters are optimized based on the improved Seagull algorithm. S4. Calculate the maximum usable capacity and health status of the series battery pack by combining the optimal parameters.

2. The method for estimating the health status of a series battery pack based on open-circuit voltage reconstruction according to claim 1, characterized in that: In step S1, the data processing procedure includes: S11. Perform mixed noise reduction processing on the original voltage and current signals, use wavelet transform to decompose high-frequency noise components, and suppress transient interference through Kalman filter; S12. A degenerate feature sequence with temporal continuity is constructed using the sliding window method on the denoised data.

3. The method for estimating the health status of a series battery pack based on open-circuit voltage reconstruction according to claim 1, characterized in that: In step S2, the temperature compensation function is: Q Ah =∑γI(I / I ref ) α-1 (t ref / T) β Δt; Among them, Q Ah To accumulate charging capacity, I ref For the reference current, T ref The reference temperature is γ, which is a coefficient in Ah, and α and β are coefficients for current and temperature, respectively.

4. The method for estimating the health status of a series battery pack based on open-circuit voltage reconstruction according to claim 1, characterized in that: In step S3, the capacity estimation model for the series battery pack is established as follows: S31, via Q start The difference in state of charge between cells is quantified to calculate and update the initial state of charge of each cell: Among them, SOC initial It is the initial SOC, Q start ΔQ represents the unusable capacity of the battery that was not charged when charging was stopped. max Q is the charging capacity from this moment until the charging ends. aging This is the actual capacity; S32, combined with Q part The SOC change during the charging process is described, and its calculation process is as follows: Among them, SOC aging It is the actual SOC, Q part This is a partial charging capacity; S33. The SOC scale of the aged battery is converted to that of the new battery to quantify the capacity difference, and the estimated value of the open-circuit voltage is solved using the OCV-SOC relationship of the new battery. The calculation process is as follows: Among them, SOC fresh It is the SOC, Q, scaled to the level of a fresh battery. fresh It is the capacity of a fresh battery. This is the estimated value of the open-circuit voltage at this time.

5. The method for estimating the health status of a series battery pack based on open-circuit voltage reconstruction according to claim 4, characterized in that: The objective function for parameter solving is constructed using the root mean square error between the measured and estimated terminal voltages. θ=[Q start ,Q aging ,R]; Among them, U i R is the actual terminal voltage of the battery, and R is the polarization resistance during the charging process.

6. The method for estimating the health status of a series battery pack based on open-circuit voltage reconstruction according to claim 5, characterized in that: In step S3, the specific process of parameter optimization includes: S34. Initialize the seagull population size, spatial dimension, maximum number of iterations, and upper and lower bounds of the optimization parameters, where chaotic mapping is used to initialize the seagull population: X n+1 =4X n (1-X n ),X∈[0,1]; S35. Each parameter updates its position according to the migration and predation behavior formulas and calculates a new fitness to obtain the optimal position. Variable A can be used to update the parameter position. The improved expression for variable A is as follows: In the seagull optimization algorithm, because seagulls have blind spots when moving, they may get trapped in local optima. Therefore, the seagull attack method has been updated: T S (t)=C S (t)+M S (t); Among them, T S (t) represents the migratory position of seagulls adapted to multi-directional spirals, C S (t) is the latest position that does not conflict with the positions of other seagulls, M S (t) is the direction of the optimal solution, P bs (t) is the position of the optimal solution, x, y, z are the three dimensions of the current position, and a is the selection update threshold; S36. Check if the current iteration count has reached the maximum value. If it has reached the set maximum iteration count, output the current best position; otherwise, continue the search.

7. The method for estimating the health status of a series battery pack based on open-circuit voltage reconstruction according to claim 6, characterized in that: In step S4, calculating the maximum usable capacity and state of health of the series-connected battery pack includes: S41. Solve for the optimal parameters of each cell in the battery pack based on minimizing the objective function, and calculate the maximum usable capacity of the series-connected battery pack using the optimal parameters: Q pack =min(Q aging，i -Q start，i ); S42. The health status of the battery pack is the ratio of the maximum usable capacity of the aged battery pack to that of the new battery pack.

Images